Audited 31 Jul 2026·Last updated 28 Oct 2026·5 citations·Tier 1·0 uses

Friction Factor Calculator

Darcy friction factor for turbulent pipe flow from the Reynolds number and relative roughness ε/D, using Haaland's explicit form of the Colebrook equation.

Friction Factor Calculator

Darcy friction factor
0.022
The dimensionless Darcy friction factor f, for use in Δp = f (L/D) ρV²/2; divide by 4 to get the Fanning factor.
Model scope
Haaland explicit approximation for the Darcy (not Fanning) factor in turbulent, fully developed, single-phase flow; Reynolds number and relative roughness are entered, and transition flow, noncircular corrections, fittings, compressibility, and non-Newtonian behavior are outside it.

Background.

Push water through a pipe and part of the pump's effort is lost to friction at the wall, showing up as a pressure drop along the run. The Darcy–Weisbach equation puts a number on it, Δp = f × (L/D) × ρV²/2, and almost all of the physics hides in one dimensionless coefficient, the Darcy friction factor f. Get f right and pipe sizing, pump selection and flow estimates follow; get it 10% wrong and every pressure drop is 10% wrong too.

In turbulent flow f depends on two things. One is the Reynolds number, Re = VD/ν, a measure of how strongly inertia dominates viscosity. The other is relative roughness ε/D, the typical height of the wall's bumps divided by the bore. At moderate Reynolds numbers a thin viscous layer next to the wall smooths over small bumps, and f falls as Re rises. At very high Reynolds numbers that layer thins until the bumps poke through, and f levels off at a value set by roughness alone.

The standard description of that behaviour is the Colebrook equation, the basis of the Moody chart, but it is implicit: f appears on both sides and has to be found by iteration. In 1983 S. E. Haaland published an explicit alternative needing a single logarithm, and textbooks put it within about 2% of Colebrook across the turbulent range — inside the uncertainty of any real pipe's roughness.

The result is the Darcy factor, four times the Fanning factor used in some chemical-engineering texts. Only Reynolds numbers of 4,000 and above are accepted. Below about 2,300 pipe flow is laminar and f = 64/Re exactly; in between, flow flickers between regimes and no formula is dependable. Fittings, non-circular ducts, compressible gas flow and non-Newtonian fluids are also outside its scope.

What is friction factor calculator?

The Darcy friction factor f is the dimensionless coefficient in the Darcy–Weisbach equation that links the frictional pressure drop in a full, circular pipe to its length, its diameter and the kinetic energy per unit volume of the flow, ρV²/2. For fully developed turbulent flow it depends only on the Reynolds number and the relative roughness of the wall, and it is the same number whatever unit system you work in.

Physically it measures how hard the wall drags on the moving fluid. The average shear stress at the wall is τw = f × ρV²/8 — equivalently the Fanning factor, f/4, multiplied by ρV²/2 — so a force balance on a length of pipe turns that wall stress directly into the Darcy–Weisbach pressure drop. Across ordinary turbulent flow the number stays small: the same formula gives about 0.0060 for a smooth pipe at a Reynolds number of 100 million, 0.0404 for a smooth pipe at 4,000, and 0.0719 for a very rough wall with ε/D = 0.05 at 100,000.

How to use this calculator.

  1. Work out the Reynolds number first: Re = V × D ÷ ν, with V the mean velocity, D the inside diameter and ν the kinematic viscosity. Water near 20 °C has ν close to 1.0 × 10⁻⁶ m²/s; air at the same temperature is about 1.5 × 10⁻⁵ m²/s, so air needs fifteen times the velocity to reach the same Re. The Reynolds number must be between 4,000 and 100 million.
  2. Look up the wall roughness ε for your pipe material — commercial steel is usually taken as about 0.045 mm — and divide it by the inside diameter in the same unit. Use the actual bore, not the nominal size: a nominal 2-inch Schedule 40 steel pipe is about 52.5 mm inside. Enter 0 to treat the pipe as perfectly smooth, a fair approximation for drawn tubing and plastic.
  3. Read the Darcy friction factor. If the pressure-drop formula you are using is written with the Fanning factor, divide by 4.
  4. Put f into Δp = f × (L/D) × ρV²/2, or h = f × (L/D) × V²/(2g) for head loss, to get the friction loss along your straight pipe run.

The formula.

1/sqrt(f_D) = -1.8 log10[((epsilon/D)/3.7)^1.11 + 6.9/Re]

1/√f = −1.8 log₁₀[(ε/D ÷ 3.7)^1.11 + 6.9/Re]. The two terms inside the bracket stand for the two ways a wall resists turbulent flow. The roughness term, (ε/D ÷ 3.7)^1.11, dominates when the bumps on the wall protrude through the viscous sublayer; the 6.9/Re term dominates in smooth pipes, where viscosity governs the layer next to the wall. Their sum is far below 1 in any real pipe, so its logarithm is negative and the right-hand side comes out positive. The constants 1.8, 3.7, 1.11 and 6.9 are Haaland's fitted values, chosen so the explicit form reproduces the implicit Colebrook equation. Squaring the right-hand side and taking the reciprocal gives f. For a fixed roughness, raising Re lowers f until the 6.9/Re term becomes negligible; at ε/D = 0.001 f settles towards 0.0197 in that fully rough limit.

Colebrook's 1939 equation, which Haaland's form imitates, joined two older laws: the smooth-pipe law of Prandtl and von Kármán, in which f depends on Re alone, and the fully rough law from Nikuradse's experiments on pipes lined with uniform sand grains, in which it depends on ε/D alone. That history explains what ε really is. It is not a measured height of the bumps on a real pipe but an equivalent sand-grain roughness — the sand size that would produce the same friction — and the values tabulated for commercial materials were inferred from friction tests.

Against an iterated Colebrook solution, Haaland's version runs about 1.3% high for a smooth pipe at Re = 4,000, about 0.9% low for a smooth pipe at Re = 100,000, and within 0.2% for rough walls with ε/D between 0.01 and 0.05. Outside the range it was fitted to, the equation stops meaning anything. The Moody chart ends at ε/D = 0.05; far beyond that a 'roughness' comparable to the bore no longer describes a pipe, and as (ε/D ÷ 3.7)^1.11 approaches 1 the logarithm heads to zero and the formula returns nonsense, so the calculator refuses a relative roughness above 0.05 and a Reynolds number above 100 million, the edges of the chart.

A worked example.

Example

Take a Reynolds number of 100,000 and a relative roughness of 0.001 — water at about 2 m/s in a 50 mm steel pipe, for instance. Roughness term: 0.001 ÷ 3.7 = 2.7027 × 10⁻⁴, which raised to the power 1.11 becomes 1.0947 × 10⁻⁴. Reynolds term: 6.9 ÷ 100,000 = 6.9 × 10⁻⁵. Their sum is 1.7847 × 10⁻⁴, whose base-10 logarithm is −3.7484. Multiplying by −1.8 gives 1/√f = 6.7472, so f = 1 ÷ 6.7472² ≈ 0.02197, which the calculator displays as 0.022. In Darcy–Weisbach terms, 100 m of that pipe loses 0.02197 × (100 ÷ 0.05) × 1000 × 2² ÷ 2 ≈ 87,900 Pa, close to 0.88 bar. Iterating the Colebrook equation for the same inputs gives 0.02217, under 1% higher. Pipe problems often arrive the other way round: the same 100 m run may lose no more than 0.5 bar, and the question is how fast the water can go. Velocity is unknown, so Re is too. Guess f = 0.022, which gives V = √(2 × 50,000 × 0.05 ÷ (0.022 × 100 × 1,000)) = 1.508 m/s and Re ≈ 75,400. Entering that Re returns f = 0.0226, revising V to 1.488 m/s; one more round barely moves it, and the answer settles at about 1.487 m/s, or 2.92 litres a second. Two passes are usually enough because f changes slowly with Re.

reynolds Number100,000
relative Roughness0.001

Frequently asked questions.

Is this the Darcy or the Fanning friction factor?
Darcy. The Fanning factor, used in some chemical-engineering texts and heat-transfer correlations, is exactly one quarter of it, so the example's 0.02197 corresponds to a Fanning factor of 0.00549. Mixing the two up is a classic way to be out by a factor of four in a pressure drop.
Why are Reynolds numbers below 4,000 refused?
Haaland's formula describes turbulent flow only. Below about 2,300 flow in a round pipe is laminar and the friction factor is exactly 64/Re, regardless of roughness. Between roughly 2,300 and 4,000 the flow switches unpredictably between laminar and turbulent, and no single friction factor is dependable there. The gap is wide: the laminar formula gives 0.032 at Re = 2,000, while a smooth pipe in turbulent flow at 4,000 has about 0.0404, and a pipe in between can sit anywhere in that range.
How do I get the relative roughness of my pipe?
Look up the absolute roughness ε for the pipe material and divide it by the inside diameter in the same unit. Commercial steel is commonly taken as 0.045 mm, so a 45 mm bore gives ε/D = 0.001. Roughness grows as pipes corrode or scale, so older lines deserve a higher value.
Which roughness values are typical for common pipe materials?
The table usually printed beside the Moody chart gives about 0.0015 mm for drawn copper or brass tubing, 0.045 mm for commercial steel, 0.15 mm for galvanized iron and 0.26 mm for cast iron, while glass and plastic are treated as smooth. These are values for new pipe. If corrosion pushed a line's ε/D from 0.001 to 0.005, f at Re = 100,000 would rise to 0.0313, 1.42 times the clean value, and the pressure drop with it.
How much does roughness change the friction factor?
A great deal in rough pipes, less near the smooth end. At Re = 100,000, ε/D = 0.0001 gives f ≈ 0.0183, 0.001 gives 0.0220 and 0.01 gives 0.0385. A perfectly smooth pipe at the same Reynolds number comes out at about 0.0178.
Why use Haaland instead of the Colebrook equation?
Colebrook has f on both sides and must be solved by iteration; Haaland gives f directly from one logarithm, which suits spreadsheets and hand checks. The price is a small error, under 1% for the worked example, usually smaller than the uncertainty in the roughness value itself. Other explicit fits exist; the Swamee–Jain equation, for instance, gives 0.02234 for the same inputs, slightly above the Colebrook value.
Does water temperature change the friction factor?
Only through the Reynolds number. Warm water is much less viscous: its kinematic viscosity falls from about 1.0 × 10⁻⁶ m²/s at 20 °C to about 0.47 × 10⁻⁶ at 60 °C, so the same 2 m/s in a 50 mm pipe gives Re ≈ 211,000 instead of 100,000. At ε/D = 0.001 that lowers f from 0.0220 to 0.0209, about 5%. In a fully rough pipe temperature hardly matters, because the Reynolds term has already dropped out.
Can I use it for air ducts or rectangular channels?
For a non-circular duct engineers commonly substitute the hydraulic diameter, 4 × area ÷ wetted perimeter, into both Re and ε/D, but no shape correction is applied here. Gas flow is fine at low speed; once density changes noticeably along the pipe, compressible-flow methods are needed.
If I double the flow, how much does the pressure drop rise?
Somewhat less than four times, unless the pipe is fully rough. Pressure drop scales with f × V², so doubling the flow quadruples V², but the higher Reynolds number trims f. At ε/D = 0.001, going from Re = 100,000 to 200,000 lowers f from 0.02197 to 0.02091, and the drop rises 3.81 times. A smooth pipe over the same range rises 3.48 times; at ε/D = 0.01 and a Reynolds number of ten million, f barely moves and the factor is 4.00.
How do I add valves, bends and other fittings to the pressure drop?
The friction factor covers straight pipe only. Each fitting adds a loss coefficient K, from the manufacturer or a hydraulics handbook, and the total becomes Δp = (f × L/D + ΣK) × ρV²/2. In the worked example's 50 mm pipe carrying water at 2 m/s, every unit of K costs 2,000 Pa, the same as about 2.28 m of straight pipe, because the equivalent length is K × D ÷ f. Pipe-sizing tables that list fittings as equivalent lengths are using that conversion.

How this page was produced

Published by
Quanta Calculator
Primary sources
5 cited below
Method
1/sqrt(f_D) = -1.8 log10[((epsilon/D)/3.7)^1.11 + 6.9/Re]
Published
Last verified

Built with AI assistance and verified by automated tests against the cited sources — every worked example on this page is computed by the same code that runs the calculator. How we build and check calculators.

In this category

Embed

Quanta Pro

Paid features are coming later.

  • All 1560 calculators remain free
  • No billing is enabled
Coming soon